Key Statistics

DATA PREPARATION

Looking for NULL values

Filling the missing values with MEAN

Merging CONTROL and TEST datasets

Chechking if the dataset has equal no. of samples

A/B testing

for the visualization above

The control campaign resulted in more impressions according to the amount spent on both campaigns.

The test campaign resulted in more searches on the website

The test campaign resulted in more searches on the website

the control campaign had more views than the test campaign

Despite low website clicks more products were added to the cart from the control campaign

There’s only a difference of around 1% in the purchases made from both ad campaigns. As the Control campaign resulted in more sales in less amount spent on marketing, the control campaign wins here!

analyzing some metrics to find which ad campaign converts more

The website clicks are higher in the test campaign, but the engagement from website clicks is higher in the control campaign. So the control campaign wins!

the control campaign had more products added to cart

Although the control campaign resulted in more sales and more products in the cart, the conversation rate of the test campaign is higher

Conclusion From the above A/B tests, we found that the control campaign resulted in more sales and engagement from the visitors. More products were viewed from the control campaign, resulting in more products in the cart and more sales. But the conversation rate of products in the cart is higher in the test campaign. The test campaign resulted in more sales according to the products viewed and added to the cart. And the control campaign results in more sales overall. So, the Test campaign can be used to market a specific product to a specific audience, and the Control campaign can be used to market multiple products to a wider audience.